By John Mullin, Enlearn CEO
Adaptive learning is clearly a popular, if overused, phrase in education technology these daysĀ ā anyone serving up digital content or curriculum is calling it adaptive. As the field widens to include more players, it is becoming clear that we at Enlearn see adaptive learning differently than most. In the prevailing adaptive model, content is finite. Publishers take their best guess at a one-size-fits-all curriculum. Ā It is then made āadaptiveā, meaning the sequence of problems or concepts can be shuffled per student based on their demonstrated understanding. āGot the question right? Hereās the next, harder problem. Got it wrong? Hereās an easier one.ā This initial approach to adaptive curriculum was a great first step, but in this model there is always an artificial cap on the content and practice available to each student.
As such, this model has several inherent limitations that weāve addressed in the Enlearn Platform. The first is this artificial constraint on content and practice. In our pilots, we found that some students need as much as six times the content as other students to achieve mastery. And that content had to be created and curated specifically to each student. It is impossible for any finite curriculum, no matter how you shuffle it, to generate the volume of content needed to create a truly personalized curriculum for each individual student.
The Enlearn Platform can now break through this constraint. Through Generative Adaptation, we can take exemplar problems from any text or curriculum and generate thousands more problems that exercise the same thought processes. Additionally, we can vary the difficulty between problems along one single dimension of complexity each time, creating an incredibly granular and virtually infinite āmapā of problems for students to traverse. For example, with six exemplar problems from one of our content partners, the platform was able to generate over 150,000 problems that cover the entire space from combining like terms to quadratic inequalities. We were able to eliminate any constraint on the content available to be specialized per student for this partner.
A second limitation of current adaptive models is relying on a studentās final answer to drive adaptivity. As we know, there may be several misconceptions or incorrect approaches to solving a problem that lead to a wrong answer. If adaptivity relies on that answer, the algorithms begin an iterative guessing game to find the problem or concept that helps a student through that roadblock. Weāve taken adaptive learning one step further at Enlearn. Our platform diagnoses the specific misconception while the student is solving the problem. We understand exactly where in the thought process students are running into trouble.Ā Knowing this, the Enlearn Platform can then deliver a curated problem, hint, or explanation shown to best address the specific misconception. Ā We no longer have to run students through a trial and error process to find the right learning progression. Enlearn can decode a studentās misconceptions or erroneous thinking and get them immediately on the right track, no guessing game (and resulting student frustration) required.
Another huge hole in most current adaptive models is that they work in isolation from, rather than through, the teacher. We recognize that teachers remain the single greatest determinant of student success at school, and our platform is designed to enhance rather than bypass the teacher. More on how we do this in our next postā¦.
